Ever since poaching became a widespread practice in Africa during the late 20th century, the African elephant has been under the constant threat of extinction. While there have been constant efforts to maintain and re-establish the dwindling population of elephants, keeping track of them has been a difficult task. As late as 2015, researchers had used aircraft and GPS-tracked cameras to survey across 18 nations and 600,000 square miles across the continent; however, these counts still had to be cross-referenced with ground-teams that surveyed nearby areas and took pictures of their own. However, in a recent development led by the University of Bath in England, this may change permanently.
New Technologies, New Opportunities
The technology that has recently emerged in several fields of science and industries has been the use of machine learning and artificial intelligence. In this case, thanks to developing satellite imaging technologies and advancements made in machine learning, counting elephants from space have become a possibility.
However, this is not a first for wildlife conservation. Satellite imagery has been used by NASA to discover a secret penguin colony, and they have also been used to collect data on whales, which are easy to spot against the blue water. What makes this project special is the variety of terrains such as woodlands, grasslands, and deserts that the program has to filter through in order to count the number of elephants accurately. Unlike the whales or penguins, which are often in a unicolor, uniformly colored environment, it is much more difficult to accomplish similar results on land.
With only 40,000 to 50,000 specimens surviving in the wild, it is more important than ever to keep an accurate record of all wild African elephants.
According to Olga Isupov, the developer of the algorithm, as well as his colleagues, “We need to know where the animals are and how many there are” in order to protect these species. The team is currently working on ways to adapt the technology to smaller animals, maybe even down to the size of mice. Unfortunately, this is not possible with current satellite camera technology, but it will most likely happen in the very near future, as the team continues to refine their current project.